Post-Mortem 7.0-Tesla Magnetic Resonance Imaging of the Hippocampus in Progressive Supranuclear Palsy with and without Cerebral Amyloid Angiopathy

: Introduction and Purpose: Cerebral amyloid angiopathy (CAA) can be observed in patients with progressive supranuclear palsy (PSP), though to a lesser degree than in Alzheimer’s disease. The present post-mortem 7.0-tesla magnetic resonance imaging (MRI) evaluates whether CAA has an inﬂuence on the degree of hippocampal atrophy (HA) and on the incidence of associated micro-infarcts (HMIs) and cortical micro-bleeds (HMBs). Material and Methods: Eight brains with PSP-CAA were compared to 20 PSP brains without CAA. In addition to the neuropathological examination, the hippocampus was evaluated on the most representative coronal section with T2 and T2*-weighted MRI sequences. The average degree of HA was determined in both groups. The incidence of HMIs and HMBs was also compared as well as the frequency of cortical micro-infarcts (CoMIs) and cortical micro-bleeds (CoMBs) in the hemispheric neocortex. Results: The neuropathological examination showed a higher incidence of lacunar infarcts in the PSP-CAA brains compared to the PSP ones. With magnetic resonance imaging (MRI), the severity of HA and the incidence of HMIs and HMBs was similar between both groups. Additionally, the frequency of CoMIs and CoMBs in the neocortex was comparable. Conclusions: The association of CAA in PSP brains has no inﬂuence on the degree of HA and on the incidence of the small cerebrovascular lesions in the hippocampus as well as in the neocortex. and the cortical micro-infarcts and micro-bleeds in the neocortex between brains with progressive supranuclear palsy associated to cerebral amyloid angiopathy (PSP-CAA) and those without this association (PSP) on magnetic resonance imaging.


Introduction
Progressive supranuclear palsy (PSP) is a sporadic disorder with tau pathology [1]. The isoforms of four-repeat tau due to splicing of exon 10 define the tau filantous aggregates [2]. Phosphorilated TDP-43 pathology can be observed in PSP, although some other studies suggest a lack of TDP-43 pathology [3].
There are only a few studies concerning the hippocampal involvement in the pathology of PSP. In one study with magnetic resonance imaging (MRI), bilateral atrophy of both hippocampi together with the severe involvement of the thalami, the pallidum, and the brainstem is described [4]. However a fluorodeoxyglucose (FDG) positron emission tomography study showed, in addition to a global decrease in several regions of the cerebral hemispheres, a relative metabolic increase in the hippocampus [5]. Cerebral amyloid angiopathy (CAA) is absent in the normal aging brain [6]. It can be associated to PSP, but does not increase the overall incidence of small cortical vascular lesions, with the exception of cortical superficial siderosis [7].
Hippocampal atrophy (HA) in PSP is much less severe than in Alzheimer's disease and frontotemporal lobar degeneration [8].
The present post-mortem 7.0-tesla MRI study investigates whether the presence of CAA in PSP brains influences the degree of HA and the incidence of hippocampal micro-infarcts (HMIs) and hippocampal micro-bleeds (HMBs). Additionally, the frequency of cortical micro-infarcts (CoMIs) and cortical micro-bleeds (CoMBs) in the cerebral neocortex is compared between PSP-CAA and PSP brains.

Material and Methods
The examined post-mortem brains consisted of 8 PSP-CAA ones and 20 PSP ones without CAA. A previously obtained informed consent of the patients or from the nearest family allowed an autopsy for diagnostic and scientific purposes. The brain tissue samples were acquired from the Lille Neuro-Bank of the Lille University, which is part of the "Centres des Resources Biologiques" and acts as an institutional review board.
The neuropathological examination of PSP and associated small cerebrovascular lesions was made according to a previously described standard procedure [9].
The clinical diagnosis of PSP was made according to criteria proposed in 2017 by the international Parkinson and Movement Society, and was classified as probable, possible, and suggestive [10,11]. However, neuropathological examination allows a definite diagnosis.
The presence of various degrees of CAA was confirmed according to the criteria of a consensus protocol, and they were graded from 0 to 3 after examining four cortical samples with β-amyloid staining [12]. Grades 1 up to 3 were retained for diagnosis in the DLB-CAA group.
The degree of HA was determined according to the Alzheimer disease (AD) classification of Scheltens in 4 grades [13,14]. Additionally, the incidence of HMIs and HMBs was evaluated as previously described for cortical hemispheric CoMIs and CoMBs [15].
A 7.0-tesla MRI Bruker BioSpin SA was used with an issuer-receiver cylinder coil of 72 mm inner diameter (Ettlingen, Germany), according to a previously described method [16]. Before the brain sampling, three to six coronal sections of a cerebral hemisphere were submitted to SPIN ECHO T2 and T2* MRI sequences. The hippocampus was evaluated on the most representative section.
Unvaried comparisons of unpaired groups were performed with the Fisher's exact test for categorical data. The non-parametric Mann-Whitney U test was used to compare continuous variables. The significance level, two-tailed, was set at ≤0.01 for significant and ≤0.001 for highly significant. Values set at ≤0.05 and >0.01 were considered as marginally significant.

Results
There was no difference between the average age at death of the PSP brains with and without CAA, with 77 (SD: 11) years in the former group and 74 (SD: 9) in the latter groups. Additionally, no significant differences in gender distribution were observed: 50% males in the PSP-CAA group and 42% in the PSP group without CAA.
There was no difference between the average age at death of the PSP brains with and without CAA, with 77 (SD: 11) years in the former group and 74 (SD: 9) in the latter group. Additionally, no significant differences in gender distribution were observed: 50% males in the PSP-CAA group and 42% in the PSP group without CAA.
The neuropathological examination revealed no statistical differences in the intensity of white matter changes and the incidence of territorial infarcts, lobar haematomas, cortical micro-infarcts, and cortical micro-bleeds in the neocortex of the cerebral hemispheres. Only lacunar infarcts were statistically more frequent in the PSP-CAA brains (Table 1). Table 1. Comparison on neuropathological examination of the average incidence of the cerebrovascular lesions between brains with progressive supranuclear palsy and cerebral amyloid angiopathy (PSP-CAA) and those without cerebral amyloid angiopathy (PSP).

Items PSP-CAA PSP p-Value
White matter changes On MRI examination, no differences in the degree of HA and in the incidence of HMIs and HMBS was observed between the PSP-CAA brains and the "pure" PSP brains (Figures 1 and 2). Additionally, no statistical differences in average frequency CoMIs and CoMBs were observed in the neocortex of either group (Table 2). Table 2. Comparison of the severity of the hippocampal atrophy and the incidence of hippocampal micro-infarcts and micro-bleeds, and the cortical micro-infarcts and micro-bleeds in the neocortex between brains with progressive supranuclear palsy associated to cerebral amyloid angiopathy (PSP-CAA) and those without this association (PSP) on magnetic resonance imaging.

2) NS
NeuroSci 2020, 1,3 and cortical micro-bleeds in the neocortex of the cerebral hemispheres. Only lacunar infarcts were statistically more frequent in the PSP-CAA brains (Table 1). Table 1. Comparison on neuropathological examination of the average incidence of the cerebrovascular lesions between brains with progressive supranuclear palsy and cerebral amyloid angiopathy (PSP-CAA) and those without cerebral amyloid angiopathy (PSP). On MRI examination, no differences in the degree of HA and in the incidence of HMIs and HMBS was observed between the PSP-CAA brains and the "pure" PSP brains (Figures 1 and 2). Additionally, no statistical differences in average frequency CoMIs and CoMBs were observed in the neocortex of either group ( Table 2). Table 2. Comparison of the severity of the hippocampal atrophy and the incidence of hippocampal micro-infarcts and micro-bleeds, and the cortical micro-infarcts and micro-bleeds in the neocortex between brains with progressive supranuclear palsy associated to cerebral amyloid angiopathy (PSP-CAA) and those without this association (PSP) on magnetic resonance imaging.

Discussion
7.0-tesla MRI allows a better detection of HMIs and HMBs than the 1.5 and 3.0 ones [17]. The present study shows that CAA has no influence on the degree of HA or on the incidence of the small hippocampal cerebrovascular lesions. These lesions are also comparable in the cerebral neocortex. On neuropathological examination, only lacunar infarcts were increased in the PSP-CAA group. No distinction could be made between them and dilated perivascular spaces [18].
PSP is possibly linked to corticobasal degeneration (CBD), and an overlap is not uncommon [2,19]. In our small series of five cases of CBD, none had associated CAA. This observation does not allow definite conclusions on the relation between PSP and CBD. However, in both diseases, the HA is linked to TPD-43 pathology [3].
AOE E4 gene is absent in pure PSP brains [20] but can be present when an Alzheimer type of pathology is associated [21].
The severity of HA and the incidence of HMIs and HMBs is also not influenced by CAA in other neurodegenerative diseases such as AD disease [22] and in Lewy body dementia [23].
It therefore appears that the hippocampus in PSP brains is protected from additional damage caused by CAA. This is similar to the observations in other neurodegenerative diseases and in contrast to the high incidence of small cerebrovascular lesions in the cerebral neocortex [24].
Further studies are needed to clarify the significance of these findings.

Discussion
7.0-tesla MRI allows a better detection of HMIs and HMBs than the 1.5 and 3.0 ones [17]. The present study shows that CAA has no influence on the degree of HA or on the incidence of the small hippocampal cerebrovascular lesions. These lesions are also comparable in the cerebral neocortex. On neuropathological examination, only lacunar infarcts were increased in the PSP-CAA group. No distinction could be made between them and dilated perivascular spaces [18].
PSP is possibly linked to corticobasal degeneration (CBD), and an overlap is not uncommon [2,19]. In our small series of five cases of CBD, none had associated CAA. This observation does not allow definite conclusions on the relation between PSP and CBD. However, in both diseases, the HA is linked to TPD-43 pathology [3].
AOE E4 gene is absent in pure PSP brains [20] but can be present when an Alzheimer type of pathology is associated [21].
The severity of HA and the incidence of HMIs and HMBs is also not influenced by CAA in other neurodegenerative diseases such as AD disease [22] and in Lewy body dementia [23].
It therefore appears that the hippocampus in PSP brains is protected from additional damage caused by CAA. This is similar to the observations in other neurodegenerative diseases and in contrast to the high incidence of small cerebrovascular lesions in the cerebral neocortex [24].
Further studies are needed to clarify the significance of these findings.

Conflicts of Interest:
The authors have nothing to declare in relation to this article.